[e2e] What should e2e protocols know about lower layers?

Joe Touch touch at ISI.EDU
Thu Oct 11 08:55:00 PDT 2001


Lloyd Wood wrote:

> On Wed, 10 Oct 2001, Joe Touch wrote:
> 
> 
>>We're also seeing a proliferation of NAT boxes.
>>
>>At some point we have to decide what is an appropriate or meaningful 
>>part of the architecture. Certainly misconfigured things aren't.
>>
> 
> but they should be tolerated.


This can't be realistic. Misconfigured things operate either 
incorrectly, with lower performance, more network load, or not at all. 
This is no different.

>>Perhaps such NATs and long links configured as broadcast should be
>>in the same bin.
>>
> 
> long links configured as broadcast?
> 
> Do you have a more specific objection to satellite broadcast links or
> the use of geostationary orbit for network traffic?


The only issue with broadcast long links is when the broadcast is either 
inefficient or gives a misleading response. Neither is true for long, 
truly broadcast shared media.

The problem is long point-to-point links that are configured as if they 
were shared media.

Joe





More information about the end2end-interest mailing list